The 10 Root Conditions Explained
Condition 1: Multiple Data Sources as Conflicting Information
Organisations maintain employee data across disparate systems (HR, payroll, performance management, recruitment) that contain conflicting information for the same data points. Different systems may show different hire dates, job titles, or reporting structures for the same employee.
Why It Matters for Predictive People Analytics: Conflicting employee records prevent accurate turnover prediction models, succession planning algorithms, and performance forecasting. Flight risk predictions require consistent employee histories—conflicts make it impossible to establish reliable baseline patterns.
Why It Matters for AI Enablement: Unresolved data conflicts prevent achieving even basic data readiness. AI models cannot distinguish between legitimate data variation and system-generated conflicts, resulting in models that learn from noise rather than signal.
Condition 2: Subjective Judgement in Data Production
Human decision-making during data collection—performance ratings, competency assessments, categorisation choices—introduces inconsistency and bias that varies across managers, departments, and time periods.
Why It Matters for Predictive People Analytics: Manager subjectivity in performance ratings creates bias that undermines high-potential identification and promotion readiness predictions. Different rating standards prevent accurate talent analytics and succession planning.
Why It Matters for AI Enablement: Uncontrolled subjectivity creates bias and noise that degrades model accuracy and fairness. AI systems trained on subjectively-collected data perpetuate and amplify existing biases, creating ethical risks in automated talent decisions.
Condition 3: Resource Limitations Affecting Data Access
Infrastructure and processing constraints restrict access to comprehensive historical datasets and limit the volume of data that can be analysed simultaneously.
Why It Matters for Predictive People Analytics: Limited historical data restricts workforce planning models and career progression pattern analysis. Attrition models require years of data—resource constraints that limit data retention directly reduce predictive accuracy.
Why It Matters for AI Enablement: Without adequate computational resources, organisations cannot process the data volumes required for robust AI training. Resource constraints force simpler, less accurate modelling approaches or abandoned AI initiatives.
Condition 4: Security and Accessibility Balance Considerations
The tension between protecting sensitive employee data and enabling analytical access creates practical constraints on what insights can be generated whilst maintaining privacy and regulatory compliance.
Why It Matters for Predictive People Analytics: Overly restrictive access prevents building accurate pay equity models, wellbeing predictions, and diversity analytics. Privacy constraints may prevent linking performance data with engagement surveys, limiting talent management insights.
Why It Matters for AI Enablement: Insufficient security creates GDPR and compliance risks that halt AI initiatives entirely. Organisations must implement privacy-preserving techniques to balance analytical capability with ethical obligations—failure prevents progression beyond basic DRL levels.
Condition 5: Diverse Coding Systems Across Functions
Different departments use varying terminology, classification systems, and taxonomies for job roles, skills, locations, and organisational structures, making cross-functional analysis problematic.
Why It Matters for Predictive People Analytics: Inconsistent job taxonomies prevent cross-functional skills analysis and internal mobility modelling. Skills ontology inconsistencies mean talent marketplaces cannot match employees to opportunities, and succession planning cannot identify candidates across divisions.
Why It Matters for AI Enablement: Inconsistent coding prevents cross-functional analysis and creates integration challenges AI cannot overcome without extensive preprocessing. Machine learning algorithms require standardised categorical variables—diverse systems delay deployment and increase costs.
Condition 6: Complex Data Representation Challenges
People data spans from highly structured formats (demographics, transactions) through semi-structured (competency frameworks) to unstructured formats (performance notes, CVs, emails, collaboration patterns), each requiring different analytical approaches.
Why It Matters for Predictive People Analytics: Advanced analytics requires integrating structured turnover data with unstructured exit interviews to understand why employees leave. Skills inference from CVs requires natural language processing. Organisations limited to structured data access only a fraction of available insights.
Why It Matters for AI Enablement: Organisations must develop capabilities across multiple AI domains—machine learning for structured data, natural language processing for text, network analysis for relationships. Single-modality AI capabilities severely constrain analytical possibilities and competitive advantage.
Condition 7: Data Volume and Processing Relationships
The relationship between data quantity and processing capability creates strategic trade-offs. Large organisations generate vast amounts of people data but often lack infrastructure to process it at scale.
Why It Matters for Predictive People Analytics: Enterprise workforce planning for 50,000+ employees requires processing millions of historical data points. Real-time flight risk monitoring needs continuous processing—insufficient power forces batch processing that delivers insights too late for intervention.
Why It Matters for AI Enablement: Insufficient processing power means models cannot be trained effectively on complete datasets, limiting AI ROI. Organisations must invest in scalable infrastructure or accept severely constrained capabilities that prevent competitive advantage.
Condition 8: Data Input Standards and User Behaviour Patterns
Formal data collection requirements often diverge from actual user behaviour. Managers skip mandatory fields, employees provide incomplete information, and shortcuts create systematic quality issues.
Why It Matters for Predictive People Analytics: Incomplete skills profiles prevent accurate talent marketplace matching and skills gap analysis. Delayed or rushed performance reviews create timing biases that undermine talent calibration models and produce unreliable recommendations.
Why It Matters for AI Enablement: The gap between intended and actual collection creates systematic quality issues that reduce model accuracy and trustworthiness. AI cannot distinguish between legitimately missing data and data missing due to user behaviour—undermining model confidence and explainability.
Condition 9: Evolving Information Requirements
Organisational needs change over time, creating misalignment between historical data collection methods and current analytical requirements. What organisations need to measure today differs from what they measured historically.
Why It Matters for Predictive People Analytics: Historical data may lack remote work patterns, gig relationships, or digital collaboration metrics now critical for hybrid workforce planning. Organisations that didn't capture skills data cannot retrospectively build skills inventories for talent marketplace initiatives.
Why It Matters for AI Enablement: Legacy data structures may not capture variables needed for modern AI applications, requiring costly retrospective enhancement, limiting model scope, or preventing deployment entirely. Rigid structures prevent agility; constantly changing structures prevent historical trend analysis.
Condition 10: System Integration and Information Architecture
The challenge of integrating distributed information systems (HR, payroll, learning, recruitment, performance, engagement, collaboration tools) whilst maintaining comprehensive analytical access and enterprise-wide data consistency.
Why It Matters for Predictive People Analytics: Enterprise analytics requires a unified employee lifecycle view from recruitment through exit. Siloed systems prevent holistic insights: recruitment disconnected from performance, learning isolated from outcomes, engagement separated from retention patterns.
Why It Matters for AI Enablement: Poor integration is the most significant barrier to achieving the highest DRL levels, preventing the unified data environment essential for enterprise-wide predictive analytics. Organisations cannot progress beyond departmental pilots to enterprise-scale AI deployment that delivers transformational ROI.
Using These Conditions
Each condition represents a potential barrier to achieving data readiness for Predictive People Analytics and AI enablement. Organisations should:
Assess which conditions currently affect their data environment
Prioritise conditions that most significantly constrain their DRL progression
Address root causes systematically rather than treating symptoms
Measure progress as conditions are resolved and DRL levels advance
Maintain ongoing attention to prevent regression across all conditions
Successfully addressing these 10 root conditions creates the data maturity foundation essential for Predictive People Analytics and AI investments to deliver expected ROI.
